Nasty Surprises

Likely you remember the Gulf War. The massive deployment of American military might in Saudi Arabia, the swift liberation of Kuwait, and the thorough drubbing of the forces of Iraq and its maniacal leader Saddam Hussein. You might also recall the stirring videos that pu orted to show Iraqi SCUD missiles being blasted out of the sky by American ground-to-air defense missiles. This was considered a pleasant 8 18 by some military experts who had long questioned the effectiveness of American antimissile weaponry.

If youre a tad practical, you may have thought that the good news coverage would send investors scurrying to buy stock in missile defense contractors. Thats what 8 18 8 do on Wall Street, they constandy change companies prospects-i.e., their earnings outlooks-and dius inevitably their stock prices. The market is always adjusting to surprises, or anticipating them, or discounting diem-the standard fare of investment news you hear or read every day. The key players, of course, in making a 8 18 , a market 8 18 , are the analysts we met in the last chapter, the people who predict what will happen barring any surprises.

Incidentally, there are analysts who look at the performance of military hardware. And despite the optimistic field reports, and the stirring videos, the actual success of the antimissiles based on a rigorous statistical study wasnt very high. The failure of the SCUDS to perform was due both to their defects and the unrealistically high performance expected from their firing crews. Time to think twice about that stock purchase? Maybe. But far more important, maybe its time to reconsider the whole market mechanism of 8 18 8, not simply from an anecdotal perspective, but from a solid statistical basis.

Paying Through the Nose for Growth

At times, no price seems too high for aggressive growth stocks or IPOs (initial public offerings). Investors repeatedly pay through the nose, and just as repeatedly get stung. Nevertheless, as chapters 4 and 5 showed, strong psychological forces compel investors to buy sizzling issues, and then prevent them from even analyzing where they went wrong.

The pattem is, not surprisingly, repeated for larger companies. Investors believe they can forecast the prospects for both exciting and unexciting stocks well into the future. They have high expectations for "best stocks," and high confidence that expectations will be met. Similarly, they have low expectations for stocks that appear to have lackluster or poor prospects, but again high confidence that their estimates will be dead on.

Companies with the best prospects, fastest growth rates, and most exciting concepts normally trade at a high price relative to eamings (P/E), cash flow (P/CF), and book value (P/BV), and invariably provide low or no dividend yields. Conversely, stocks with poor outlooks trade at low P/Es, price-to-cash flow, or price-to-book value, and usually have higher dividend yields. (For information on where to find price-earnings, price-to-book value, price-to-cash flow ratios and dividend yields of companies, please tum to pages 202-203, chapter 9.)

Often, the disparity between what investors will pay for a favored stock and one blacklisted is immense. People in early 1996, for example, were willing to pay 23 times as much for each dollar of eamings of Netscape, the Intemet wunderlcind, as for Fannie Mae (Federal National Mortgage Association), which incidentally is no slouch in expanding income, having increased it at more than 20% annually over the previous twenty years. But investors will pay these price differentials, which can be enormous, because of their confidence in their ability to pinpoint the future. Lets look at what happens when-su rise!-their forecasts miss the mark.

Eamings surprises have a consistent and predictable effect on stoclc prices. More to the point, their impact on stoclcs that people Uke is dramatically different from their impact on stocks they dont Uke. Understanding the nature of surprises provides a high-probability method of beating the market.

What the Record Shows

Next we set a yardstick to gauge the result of market 8 5 . The surprises are measured against the analyst consensus forecast, the average

The Many Faces of Surprise

To find out how stocks react when analysts err, I did a number of studies in collaboration with Nelson Woodard, Eric Lufkin, Michael Berry, and Mitchell Stem for the 24 years ending with 1996. To be consistent, we worked with exactly the same analysts consensus forecasts that were used to calculate analysts errors in the last chapter.

We wanted to measure a number of factors important to investors. First, what do analyst forecasting errors do to stock prices? Second- and as important-do eamings 5 18 5 have the same impact on favored as on unfavored stocks? Stocks trading in outer space are there because of analysts confidence in their future-possibly mixed with a wee dose of overoptimism. Did they react the same way to eamings surprises as stocks that are in the investor doghouse? Third, we wished to examine just how accurate investors expect analysts forecasts to be. To help resolve this, we measured how even tiny 8 18 8 affect a stocks price by considering any amount over one penny a share a 5 18 .

To answer the three questions, we analyzed the stocks stricdy according to how exciting or dull investors believed their prospects were, using three different value measures: price-eamings ratio, price-to-cash flow, and price-to-book value. The higher the three ratios, the more enticing the stock is to an investor, and the more he or she is willing to pay. Conversely, the lower the three ratios, the more unpopular the stock. We divided the stocks in all of the quarters in our 1973-1996 study into three groups stricdy by how they ranked by each of these three value measures. The 20% of stocks that had the highest P/Es, for example, were placed in the top P/E group (called a quintile in statistical jargon), the next 60% in the middle group, the lowest 20% in the bottom group. We did this the same way for all three value measures. The portfolios were reassembled on this basis for every quarter in the study. We then calculated the effect of on each group of stocks beginning in the second quarter of 1973 and ending in the fourth quarter of 1996.

The study used the 1,500 largest companies in the Compustat database with fiscal years ending in March, June, September, and December." Approximately 750-1,000 large companies were used in each of the 95 quarters of the study.